Research on Digital 3D Reconstruction of Cultural Relics Based on Convolutional Enhanced Neural Radiance Field

Relics are important carriers of human history and culture, bearing rich historical information and cultural heritage. How to accurately reconstruct and preserve cultural relics is an urgent problem that needs to be addressed. Digital modeling effectively retains the visual characteristics of relics, making its application in reconstruction highly significant. Traditional visual reconstruction methods require multiple viewpoint images, which incur high time costs and lack efficiency. To tackle this issue, this paper uses camera captures of various cultural relics to build a dataset and employs a new neural radiance field - convolutional enhanced radiance field for the digital reconstruction of these relics. Compared to the original radiance field, this method achieves better reconstruction results, with clearer details of the artifacts and avoids common holes found in traditional mesh reconstruction. This provides a new perspective for the reconstruction of cultural relics.